Chapter in N. Gilbert (ed) (2008) Researching Social Life, 3rd Edition, London: Sage.

21
NARRATIVE ANALYSIS

Sarah Earthy and Ann Cronin

Key points
21.1 Introduction
21.2 What is narrative analysis?
21.2.1 The traditional status of qualitative data
21.2.2 Narrative analysis
21.2.3 Definitions
21.2.4 Example: narratives of chronic illness
21.2.5 Example: stories of homelessness
21.3 The social production of the story
21.3.1 Stories as symbolic interactions
21.3.2 Stories as political process
21.3.3 Representation
21.4 Using a narrative approach
21.4.1 Planned and unplanned generation of narrative data
21.4.2 Oral history and life-story research
21.4.3 Multiple interviews
21.4.4 Asking narrative questions
21.4.5 Textual sources of narrative data
21.5 Analysing data as narrative
21.5.1 Unit of analysis: categorical versus holistic
21.5.2 Focus of analysis: content versus form
21.5.3 Combinations of unit and focus of analysis
21.5.4 Example: a story of hardship
21.5.5 Validity and trustworthiness
21.6 Summary
Discussion questions
Project
Resources
KEY POINTS
 • Story telling is a natural part of social life – we all tell stories in different situations and different
ways.

• Stories are not created in a social vacuum but are moulded by a social context and interactions
between the story-teller and audiences.

• Social researchers are interested in the kinds of stories people tell, how stories are presented and why
they are told.

• There are different approaches to narrative analysis but a common starting-point is the rejection of
traditional realist assumptions about research data in favour of a social constructionist position.

21.1 Introduction

Social research has a long tradition of using qualitative methods to gain an insight into people’s lives and to
understand the meaning that people attach to their lived experience. Chapter 12 examines how focus groups
can be used for this purpose, while Chapter 13 provides an introduction to qualitative interviewing. This
chapter builds on these previous discussions by focusing on qualitative data as narrative and the
implications for both data collection and analysis. In doing so, we explore how narrative analysis enables us
to explore the socially constructed nature of the research process and the role ‘stories’ play in the
construction of identity.
The chapter is divided into six sections. The first section provides an overview of narrative analysis
and its use in social research. The second section examines the social production of the story. The third
section explores how to use a narrative approach in qualitative interviewing while the fourth section focuses
on data analysis. The fifth and final section provides a summary of the main points discussed in the chapter
followed by discussion question, an outline of a student project and suggestions for further research. Finally,
while the main focus in this chapter is on data collected through individual interviews, it should be borne in
mind that narrative analysis can be as useful in the analysis of focus group and documentary data.

21.2 What is narrative analysis?

This section begins by examining the traditional status accorded qualitative data before moving to consider
how narrative analysis challenges this. It then explores what is meant by narrative analysis and discusses
some of the reasons why social researchers might choose this method to collect and analyse qualitative data.

21.2.1 The Traditional Status of Qualitative Data

The accounts people tell us about their lives form a fundamental part of social inquiry; that is, as social
researchers we gather accounts about people’s lives, which we subsequently use to produce our own
accounts of issues of concern to us. For example, an interest in exploring homelessness would probably
involve interviewing people who have experience of living on the streets and homeless shelters about their
lives. The data collected would subsequently be analysed, possibly using thematic analysis (see Chapter 13)
and the findings written up – perhaps for an undergraduate dissertation, a public report or even for
publication in a journal or book.
However, for the purposes of this chapter the important issue to consider is the epistemological status
accorded data collected in this manner. Traditionally it has been assumed that an individual account can be
regarded as representational of a real life. In other words, the social researcher can, through appropriate use
of the appropriate method, discover the ‘real subject who is present in the world’ (Denzin, 1989b: 14). Thus,
in the case cited above, the individual accounts of homelessness would be regarded as accurate
representations of the participants’ experiences of homelessness.
This understanding of data is most often associated with a methodological realist paradigm. Within this
paradigm, the account as a subject of social scientific interest in and of itself is overlooked in preference to
the information or understanding it may yield on a given topic. In contrast, narrative analysis, located within
a social constructionist paradigm, first of all challenges this realist position and, second, offers an
alternative approach to the understanding of both the production and the analysis of qualitative data.

21.2.2 Narrative Analysis

Although narrative analysis has its origins in literary theory and is closely associated with media and cultural
studies, social scientists have become interested in using it as a means to gain greater understanding of the
social world and the production of data (for example, Bruner, 1986; Denzin, 1989b; Geertz, 1975; Riessman,
1993; Rosenweld and Ochburg, 1992). Perhaps reflecting its multidisciplinary use, the term narrative
analysis can refer to a variety of different approaches to data collection and analysis, including biography,
autobiography, life history, oral history, autoethnography, life narrative and the sociology of story telling.
We take a closer look at these different methods of data collection in Section 21.4. While not dismissing the
differences between these approaches, all share a common aim, namely to explore the different ways in
which both the production and analysis of qualitative data can be understood as processes whereby different
groups of people engage in ‘story telling’ and in doing so produce narrative accounts of their lives. As a
result of this, those using narrative analysis prefer to use the terms ‘narrative’ or ‘story’ rather than
‘account’; this convention is used throughout this chapter (see Section 21.2.3 for definitions).
Rosenweld and Ochberg suggest that narrative analysis disrupts the traditional social scientific analysis,
which has realist assumptions and a focus on information collection. Instead the focus shifts to look at the
very construction of narratives and likewise the role they play in the social construction of identity.
‘Personal stories are not merely a way of telling someone (or oneself) about one’s life; they are the means by
which identities may be fashioned’ (Rosenweld and Ochburg, 1992: 1).
This interest in narratives is a part of a wider move within social sciences towards a more ‘interpretive
turn’ (Rabinow and Sullivan, 1979). This can partly be explained by a growing mistrust in the empiricist
agenda, part of which urges the researcher to adopt a more reflexive approach. As discussed below,
inclusion of the researcher in the production of data raises further doubts about the traditional claims that our
research findings can ever represent the ‘truth’ of people’s lives. The use of a narrative analysis approach
with its focus on the social construction of the story, means that uncovering the ‘truth’ no longer becomes
the object of analysis; there has been a move away from the ‘what’ to the ‘how’. This in turn deconstructs
the realist position that assumes that life stories can be regarded as ‘mirrors of life events’ (Rosenweld and
Ochburg, 1992). Before looking more closely at narrative analysis we want to raise the issue of terminology
and provide a couple of examples of narratives and stories.

21.2.3 Definitions

Account: a general term for the overall report or description given by an interviewee during a research
interview. An account may include a variety of different forms of talk and represents the interviewee’s
perceptions, understanding and experiences of the issue(s) being researched.

Narrative: a term widely used in social theory and social research to describe either: (a) a tale or story; or
(b) a form of talk or writing that aims to tell a story and may be structured according to classical ideas of
plot.

Narrative analysis: an approach taken to interview data that is concerned with understanding how and
why people talk about their lives as a story or a series of stories. This inevitably includes issues of identity
and the interaction between the narrator and audience(s).

Story: the description of an event or series of events in a manner that conveys meaning as well as factual
information. Traditional stories or myths serve a number of purposes including entertainment, instruction
and the formation of a collective worldview. When research participants tell a story or a series of stories, the
researcher will want to consider what purpose the story serves and why the interviewee has chosen to
present their account in this way.

21.2.4 Example: Narratives of Chronic Illness

An understanding of the term narrative can be found from the following example which involves
sociologists who are interested in exploring people’s experience of health and illness. Sociologists who have
interviewed people with a chronic illness such as rheumatoid arthritis have noted that these conversations
often take the form of a narrative of both the illness (how it began, possible causes, interactions with health
professionals, significant developments, day- to-day living, changed priorities, and so on) and the impact of
illness and disability on the person’s social roles and hence sense of worth (Frank, 1995; Williams, 1984).
Bury (2001) has distinguished between narratives of chronic illness that try to make sense of the experience
of illness (‘contingent narratives’) and those that address a changed self (‘moral narratives’) but in practice
these may be combined. These insights into the purpose of narratives told by interviewees can be applied to
many other interview topics, particularly where the experience being described is problematic, unresolved or
associated with stigma.

21.2.5 Example: Stories of Homelessness

Likewise, a study conducted with people who have experience of homelessness illustrates the way in which
people tell stories about their lives and the meanings they attach to them. This small-scale study formed part
of a larger mixed method study (see Chapter 7) exploring the experiential nature of vulnerability among
different groups of people living in the same location.
The original intention had been to run two focus group discussions and use the findings to inform the
development of an individual interview schedule for use with a further five participants. At this stage a
decision to use narrative analysis had not been taken. However, after running the first group it was apparent
that the members of the group were not simply supplying information on homelessness but were telling us
stories about their experiences and in the process ‘doing’ identity work. Homelessness carries with it social
stigma and thus the participants used the focus group as an opportunity to provide an alternative account of
homelessness that explored among other things causal factors, structural factors and issues of personal
agency.
Furthermore, the story telling was occurring at two levels: first, the individual stories of homelessness
and, second, a collective story of homelessness. This second level of story telling occurred in several ways,
for example, the participants might validate or support the stories told by other members of the groups, or
they would add to the story of one participant by offering their own story on a similar subject. Recognition
of the participants’ desire to tell individual stories led to a decision not to run a further focus group but to
extend the number of individual interviews from five to eight and to use a narrative analysis approach both
to collect and to analyse the data.

21.3 The social production of the story

Although narrative and story have slightly different meanings, the terms are often used interchangeably.
Furthermore, as illustrated below, the use of either ‘narrative’ or ‘story’ can also signify different
disciplinary conventions and interests. Regardless of these differences, the use of these terms is not meant to
imply that people produce fictional accounts of their lives, rather it is to draw attention to three interrelated
issues: the social production of accounts; some of the work performed by the use of stories or narratives; and
finally, the narrative-like qualities that are often present in personal accounts of life.
As discussed above, an awareness of the ‘narrative potential’ in qualitative accounts will lead the social
researcher to attend to its social production. Indeed, this in itself is a valid reason for choosing to use
narrative analysis. However, narrative analysis is not a homogenous entity and within its broad remit there
are a variety of different models available to help make us aware of the social production of the story.
Models closely aligned with literary theory or cultural and media studies, for example, concentrate on the
narrative structures present in stories. This might involve a focus on identifying the story’s plot, setting and
characterisation, which will in turn help draw attention to the way in which a story is constructed and
developed through the selective inclusion (and hence exclusion) of past events.
The sociologist Ken Plummer (1995), while retaining an interest in narrative structure, examines stories
as both symbolic interactions and political process. His interest lies in producing a sociology of story
telling concerned with: ‘[t]he social role of stories: the ways they are produced, the ways they are read, the
work they perform in the wider social order, how they change, and their role in the political process’ (ibid.:
19). In both symbolic interaction and political process we will see the importance of representation.

21.3.1 Stories as Symbolic Interactions

Symbolic interactionism suggests that all human behaviour is social, involving social interaction and the
development of shared meaning. Plummer locates the production of stories firmly within this framework and
examines the way in which the telling of stories is a central part of this symbolic interaction. This enables
stories to be viewed as joint actions involving three groups of people: the producers, the coaxers and the
consumers. This shifts the emphasis away from seeing a story as representative of an individual life to a
focus on the social production and consumption of the story.
The first group is composed of the producers of stories, those people who tell the stories of their lives.
This might be the participants of a chat show programme, the published autobiographical ‘coming out
stories’ told by lesbian and gay men; or the participants of a research project. In each of these cases (and
many others) it is important to be aware that the story is simply a selective reconstruction or version of a
life; it is not the life itself. This is the first step away from regarding the story as representative of a life, an
event or experience and seeing it as a creation in itself. This approach can help us to identify issues or events
that are significant to the individual telling their story.
The second group of people consists of the ‘coachers’ or ‘coaxers’ who play a fundamental role in the
production of the story. Plummer points out this can include an array of people: from the chat show host to
the social researcher. These people coax, persuade, and provide a forum for people to become story-tellers.
The role is not a passive one; the researcher is actively involved in the production of the story. An obvious
example is the story that results from an interview: the questions asked limit and shape the story told. This
increases the distance between the lived life and the story told about that life. The researcher’s role is
explored in detail in Section 21.4.
While the teller and the coaxer are involved in the production of a story, of equal importance in this
model of stories as joint actions are the consumers of the stories. Furthermore, the consumers are involved in
the active consumption of these stories. Just as those in media studies (for example Bobo, 1995; Gammon
and Marshment, 1988) have highlighted the interpretative and active role of the television watcher or
filmgoer, so it is with those who consume stories. Plummer argues that this means that any analysis of
stories must pay attention to the social location of the consumers. If we accept the idea that individual
consumers may construct different meanings from the same story, this shows the weakness of the link
between the life and the story told about that life. However, this analysis of the role of the consumer would
not be complete without an understanding that the producers – the tellers and coaxers – are also consumers
of other stories.

21.3.2 Stories as Political Process

To claim that stories can be understood as political process alerts us to the power mechanisms or structures
that permit certain stories to be told while silencing others. An example drawn from Plummer’s work on the
telling of sexual stories will help to illustrate this point.
Among the stories that Plummer examines are those told by women and men who identify as lesbian or
gay. Plummer observes that, over the past one hundred years of Western culture, at different times talking
about being lesbian or gay would have (and has) resulted in different consequences. Thus for much of this
century the personal stories of being lesbian and gay have either been silenced or told in secret. It is only
since the 1970s and the growth of the lesbian and gay movement that these stories have started to be told in
public. This has changed the stories from being about personal shame and pain to ones of pride and strength.
This is a clear demonstration of the practice of power, both its repressive and productive qualities.
This provides us with another reason why social scientists might be interested in adopting a narrative
approach within their research: it can be particularly useful with either marginalised groups or people who
have a discreditable or stigmatised identity.

21.3.3 Representation

From the discussion so far it will be apparent that one of the issues that narrative analysis is interested in is
the issue of representation. It is of paramount importance in Plummer’s sociology of story-telling; likewise
it is an issue that Riessman (1993) addresses in her discussion of narrative analysis and the role of the
researcher. Riessman argues that representation is ambiguous and hence open to multiple interpretations,
and identifies five levels of representation that are present in the collection and analysis of data: attending,
telling, transcribing, analysing and reading. At each level the researcher is engaged in a process of
interpretation of a life to which they have no access:

Meaning is ambiguous because it arises out of a process of interaction between people: self,
teller, listener and recorder, analyst and reader. Although the goal may be to tell the whole
truth, our narratives about others’ narratives are our world creations … Meaning is fluid and
contextual, not fixed and universal. All we have is talk and texts that represent reality
 partially, selectively, and imperfectly. (Riessman, 1993: 15)

Hence narrative analysis, whether it is undertaken using models more commonly associated with literary
theory, or using Plummer’s approach, offers the social researcher a new way of looking at and understanding
accounts. Instead of viewing an account as simply representative of an individual life, it directs attention to
the ‘joint actions’ involved in the production of the story. Moreover, it indicates the way in which the
meaning of the story and hence its consequences are always dependent on first of all the social location of
those involved in the production and consumption of the story and, second, the wider social context in which
the story is told. Plummer’s examination of the important role of ‘communities’ who hear and receive the
story, again highlights the way in which different communities over time will receive the story in different
ways. This helps us to understand the way in which stories can be understood as political processes that
involve power relations.

21.4 Using a narrative approach

A narrative approach is concerned not only with the story-telling components or characteristics of an
account, but also with the social interactions between interviewer and interviewee that encourage and
influence the way that an account is presented. To adopt a narrative approach is to choose to understand and
analyse interview or other data from that perspective rather than, for example, focusing solely on the content
of what interviewees are saying or the conversational forms and rules that underlie the interaction (e.g.
conversation analysis, described in Chapter 22).
The same section of interview data can be analysed in different ways. For example, an interviewee might
describe a traumatic event involving someone close to them, such as a delay in seeking help for what turned
out to be a serious illness. A thematic approach to such data would involve comparing accounts from a
sample of interviewees with similar experiences to explore the reasons why those involved either sought or
did not seek advice or the nature of encounters with health professionals. An ethnomethodological approach
might concentrate on the language used and how conversational devices add emphasis, enhance credibility,
deflect possible criticism, or convey the researcher’s sympathy. A narrative approach will take account of
both the content and the form of the interviewee’s account and interactions with the interviewer. More
importantly, a narrative approach will ask questions such as ‘Why is the interviewee narrating this incident
in this particular way?’, ‘What is the purpose of the story?’, ‘Why does it occur at this point in the
conversation?’, ‘How have the researcher or others present influenced the narration?’, or ‘How does this
excerpt fit with other parts of the interviewee’s life story as narrated during the interview as a whole?’.

21.4.1 Planned and Unplanned Generation of Narrative Data

Some social researchers adopt a narrative analysis approach from the outset of their research and design
their interview schedule accordingly. Other researchers do not set out to collect narratives or stories, but
instead discover the presence of story-like qualities in an interviewee’s account during the analysis stage of a
study. An example of this last position was given above in the discussion of the study with people who have
experience of homelessness.
These two styles of interviewing are not as different in practice as this distinction implies. Even in the
case of apparently spontaneous narratives, a close examination of the transcript will reveal that the framing
of questions and signals of interest and encouragement from the researcher (non-verbal as well as verbal)
precede story-telling and shape how the narrative is presented. On another occasion with different people
present, the same event or life story might be related in a different way. This is what is meant when
narratives are described as being ‘co-authored’. As we develop our interview questions, we need to be aware
that we are occupying the role described by Plummer (1995) as that of a ‘coaxer’ who may either facilitate
or inhibit the act of narration.

The following interaction provides an example of how the researcher’s likely interest
may be tested before a story is told:

RESEARCHER: Whereas here seems better...?

SUE: It’s not too bad. It has its moments.

RESEARCHER: Does it?

SUE: When we first moved in, it was terrible, very rough and
ready. Moved here September 25th ‘92.

RESEARCHER: What sorts of things were happening?

SUE: Fights ...

Since story-telling is a natural part of conversation, it is not surprising that research methods focused on
accessing an ‘emic’ perspective (i.e. an insider’s understanding of a social situation or experience) will
allow and even encourage the telling of stories. Particular topics – for example, those relating to difficult
circumstances or issues closely connected with a person’s sense of identity and self-worth – are likely to
involve story-telling and, as we have seen, certain types of question and interactions between the
interviewee and ‘audience’ may give tacit approval for a narrative to unfold.
However, as Riessman notes (1993: 56), narratives also occur when the line of questioning seems
uninviting. The resulting narrative may be more disjointed than those encountered in narrative-orientated
interviews but are nevertheless still present because the interviewee has a story to tell. It is then up to the
interviewer to choose whether to include considerations of narrative in the analysis.

21.4.2 Oral History and Life-Story Research

There are a number of research topics that lend themselves to narrative enquiry. These are often experiences
or events shared by a number of individuals that tell us something about the nature of society as well as
being significant to those involved.
Oral history aims to explore what it was like to live in a past era and to capture and preserve the
memories of a cohort while they are still alive. An example of this is Paul Thompson’s interviews with
people born during the Edwardian era in Britain (Thompson, 1992; 2000). Many public libraries and some
museums have local history collections in which interviews with older residents form a core resource and
provide rich information about aspects of day-to-day life such as childhood, family, work, religion,
shopping, transport and leisure as well as insights into social values and meaning. Such research aims to
record the experiences of ‘ordinary’ people and thereby give them a collective voice; an objective shared by
many types of qualitative social research.
Life-story research more generally is concerned with the link between personal biography and social
processes, past and present. These might involve the transformation of a place, an industry or a way of life.
Chronology is a key element in how we naturally tell stories and it is equally important to social researchers
mapping personal and social change. This is evident in the enduring usefulness of the concept of ‘career’ in
sociological research (Becker, 1961; Thomas, 2003). A similar concept is present in life-story interviews
that focus on how people are socialised into a particular occupation or social class (‘becoming’ a doctor, a
factory worker, a drug dealer, a member of an elite) or adapt to a radical change in circumstances (such as
parenthood or illness). Other types of life-story interview concern broader aspects of identity including the
enduring effects of trauma and stigma. The oral history section of the National Sound Archive at the British
Library in London (http://www. tiny.cc/7yqYE), for example, includes collections of recorded interviews
with Jewish survivors of the Holocaust and their children, and with people living with HIV and AIDS.
Increasingly, research-funding bodies are encouraging researchers to archive their recordings so that these
are available for secondary analysis by other researchers. While this has obvious practical merits, it does
have ethical implications particularly in respect of informed consent (see Chapter 13).

21.4.3 Multiple Interviews

A characteristic common to both oral history and life-story interviews is that they tend to last longer than
other forms of in-depth interviews and may involve several sessions with the same person. These might take
place over a couple of weeks or several years. For example, Robert Bogdan’s ground-breaking
autobiography of a transsexual, Jane Fry (Bogdan, 1974) and Kathy Charmaz’s study of the relationship
between self and time in chronic illness (Charmaz, 1991) are both based on multiple interviews with the
same individuals conducted over an extended time period.
Interviewing a research participant on more than one occasion can have several advantages:
• It may assist the development of trust and rapport between the researcher and interviewee.
• It may be less exhausting for both parties, particularly in comparison with a single attempt to capture
 a person’s life story.
• For interviewees who are unwell or who find aspects of the conversation distressing, the possibility
of ending the interview knowing that the conversation can continue on another day may be
particularly valuable.
• The gap between interviews provides an opportunity for both the interviewee and researcher to
reflect.
• Aspects discussed in one interview can be clarified and explored in greater depth in a subsequent
conversation.

The biographic narrative interpretive method (Wengraf, 2001) provides an example of a narrative approach
that uses three separate interviews with different formats and purposes. In the first interview, the interviewee
is invited to tell the story of their life. In the second interview, the interviewee is asked to talk more about
parts of their life discussed in the first interview and encouraged to tell further stories. On both occasions,
the researcher is concerned to minimise their influence on how the narrative is related. In the final interview,
the researcher takes the lead and asks prepared questions based on the emerging analysis. The data generated
in the three interviews is analysed in two distinct ways: as an objective account of a ‘lived-life’ and as a
subjective ‘told-story’. These separate stages of analyses are then brought together to explore why the
person has chosen to narrate their life in the particular way that they have. The central idea is that the
narrative presented in the interviews not only records the most significant events and experiences in the
person’s life and does so in greater detail than would be generated by other methods but also reveals the
interviewee’s emotions, values and beliefs during the act of narration.
While the staged process of interviewing involved in the biographic narrative interpretive method may
not be feasible for small projects with limited time, it is worth considering whether two interviews with the
same individual might produce richer and more insightful data than a single interview, even if it requires a
reduction in the size of the sample. Real rapport is not always easily established, especially when the subject
matter is sensitive. Jocelyn Cornwell, in her community study of health and illness in East London
(Cornwell, 1984), described the difference between ‘public accounts’ – those that reflected collective norms
and presented the community in a positive way – and ‘private accounts’ that disclosed information about
more difficult subjects such as domestic violence. For some research topics, multiple interviews may offer a
way of crossing from one type of account to the other.

21.4.4 Asking Narrative Questions

In both ordinary conversation and research interviews, we often ask questions that relate to chronology and
time. This is particularly true when we are trying to get to know someone, we are asking about an
experience that is unfamiliar to us, or wanting to explore similarities with our own experience (or the
experience of other interviewees). Examples of interview questions that specifically refer to time might
include: ‘When did you first…?’; ‘Starting at the beginning, tell me about…’; ‘Thinking back over the last
week…’; ‘Looking ahead...’; ‘Comparing your experiences with those of children today ...’ Similarly, when
we tell a story, we tend to structure it in a way that makes sense as a chronological sequence. We will return
to the issue of narrative structures later in this chapter but in the meantime we should note that these kinds of
questions lead the interviewee to believe that chronological order is important to us and this may influence
how their account is presented.
Most researchers who conduct interviews that are pre-designed to invite narratives recommend the use of
a combination of very broad questions (‘tell me about your life’) with prompts that encourage the relating of
specific examples (‘tell me what happened’; ‘have you ever experienced...?’; ‘can you think of a time when
...?’). Such questions may differ from those used in other types of in-depth interview if they point clearly to
the type of response wanted; that is to say a narrative or story rather than an opinion or interpretation. Even
closed questions may produce a storied response if the subject matter lends itself to story-telling or the
interviewee has a story that needs telling.
Thompson (2000: 309–23) suggests one way of structuring oral history interviews is to start with family
members (grandparents and their generation, parents, siblings and other relatives) before asking about each
life stage and related experiences (for example, memories of childhood might invoke descriptions of family
life, neighbourhood and school). Comparisons with contemporary life and perceptions of change will be
recurring themes throughout the conversation.
It is also considered good interviewing practice for the researcher’s involvement to be minimal so that
the interviewee’s narrative can flow. The interview transcript will provide a guide to the extent and content
of the researcher’s verbal contributions but not intonation or other sounds, facial expressions or body
language. These can be equally influential and are much harder to quantify except by means of notes taken
shortly after the interview and during the process of transcription.

21.4.5 Textual Sources of Narrative Data

We have mainly focused in this chapter on story-telling in the context of in-depth interviews. However, a
wide range of sources can provide access to narrative data that is already in the form of text. These include
autobiographies, biographies, newspapers and magazines, Internet/blogs and discussion forums, and
fiction. Although such data have not been generated through interaction with a researcher (except perhaps in
a discussion forum), the three roles identified by Plummer of producer, coaxer and audience continue to be
relevant.
One advantage of textual sources of narrative data is that it may be feasible to include a much larger
sample size and to use random methods of selecting cases for study. Narrative analysis can then more easily
be combined with quantitative methods such as content analysis.
 21.5 Analysing data as narrative

Having outlined some of the theoretical ideas that underpin narrative analysis and the research situations in
which narratives and story-telling may occur, this section considers methodological issues involved in
narrative analysis. The first step is to decide what aspect of narrative to explore and how the data will help to
do this. These considerations will be influenced by our ontological and epistemological standpoints (see
Chapters 1 and 2).

21.5.1 Unit of Analysis: Categorical Versus Holistic

Lieblich et al. (1998) have identified two intersecting dimensions of approaches to narrative analysis. The
first dimension concerns the unit of analysis and whether this is a category (for example, a particular type of
event or experience) or the narrative as a whole. Categorical approaches to narrative analysis compare all
references to the selected phenomenon within one interview or across several interviews, while holistic
approaches seek to understand how a particular section of text is part of a life story narrated during the
course of a single interview or several interviews with the same individual. As Lieblich and her colleagues
describe, categorical analysis tends to be used when the research is concerned with an experience that is
shared by a group of people (for example, the process of migration) whereas holistic analysis will explore
significance and change in the context of one person’s life (for example, the effects of migration on
identity).

21.5.2 Focus of Analysis: Content Versus Form

The second dimension of difference is between analytical approaches that are concerned with the content of
a story/narrative and those that focus on the way it is structured (its form). The content of the narrative may
include the surface content (what happened?, who was present?, how did different parties react?) and the
underlying or latent content (what were the motives or intentions of participants?, what might particular
items symbolise for the narrator or others?, what is the meaning and importance of this story for the
narrator?). In contrast, if we are interested in the form of the narrative, we concentrate on aspects such as
how the plot is structured, the sequence of events, and the language used. For example, particular words or
phrases may have the effect of making the story seem more convincing or protecting the narrator from being
criticised.

21.5.3 Combinations of Unit and Focus of Analysis

Different choices about the unit and focus of analysis will produce one of four possible combinations:
holistic-content, holistic-form, categorical-content and categorical-form (Lieblich et al., 1998: 13). Labov’s
interest in the structural elements of a core narrative (Labov, 1972), discussed below, is an example of a
holistic-form combination. Other studies are more difficult to categorise. For example, Riessman’s research
on narratives of divorce (Riessman, 1990; 1993) explores the narrative form and language used by
interviewees and the political contexts in which such narratives arise.
An example from our own research illustrates how these choices may be influenced as much by the
nature of the data as the research questions.

21.5.4 Example: A Story of Hardship

The following story about the shame of poverty was part of the interview data generated during a study of
contemporary community. The research comprised of ethnographic observations in two localities and 49 in-
depth interviews involving 69 participants. Although there was nothing in the interview schedule that
specifically invited stories, it became clear during fieldwork and analysis that there was something about the
subject of place and its significance for people’s identity that produced stories that frequently concerned the
past, despite a research focus on the present.

The only time I felt poor and I did feel this. Because the war had started in 1939 and I
was thirteen and the school had left to be evacuated they allowed me to leave school
at thirteen although the school age was fourteen. But there was no school to go to.
And my mother (LAUGHS) went and got, found me a job at Smith’s, the tailors. Can
you imagine going to a top quality shop as poor as a church mouse? Oh, and she was
told that the uniform, you had to provide the uniform for shop work in those days.
And I was told the uniform was black. So Lillian arrives at Smith’s, a shop where all
the people from Park Road used to get their clothes on appro and send them back if
they didn’t like ‘em. … And Lillian arrives from X in her black crêpe dress cut down.
Uneven hem because the crêpe in those days wasn’t crêpe as it is now and in this shop
with all the ladies or all the assistants that were well trained assistants because you
had to go forward and say ‘Can I help you madam?’ in those days. And I’ve never felt
so embarrassed in all my life cos they made me feel so (pause) I dunno. But my
mother, because of the area we come from, didn’t see anything in this. She didn’t see
what she’d done to a young girl who, erm, (pause). Cos, er, that’s how life was in
those days. (Lillian, aged 75)

This story can be analysed in a number of ways. First, if our unit of analysis is categorical and our focus is
the content rather than form of the story, then we will begin by looking at the surface and latent content of
this story before looking for similar stories elsewhere in the dataset.
Lillian’s story of her first day of work fits with Bury’s description of a ‘moral narrative type’ (Bury,
2001). The act of walking into a ‘top quality’ shop wearing a dress with an uneven hem fundamentally
challenged Lillian’s identity, which until that point had been within a social environment in which her
family’s economic circumstances were unremarkable. The cut-down crêpe dress, which was presumably
someone’s best, was inadequate for a shop devoted to clothes and in which all the other assistants were to
Lillian’s eyes ‘ladies’. She cannot find words to describe how the shop assistants made her feel or the
personal impact of this experience, including her mother’s inability to comprehend its significance. The final
sentence of the extract ‘that’s how life was in those days’ provides a partial resolution by emphasising the
gulf between then and now. Nevertheless the fact of recounting the story more than sixty years after it
occurred suggests that the sense of shame remains present in Lillian’s view of herself.
When we look for other examples of this kind of story in the dataset, it becomes clear that two types of
story occurred frequently in the accounts of older interviewees such as Lillian. These were stories of a
golden past and stories of material hardship from the narrator’s childhood or youth. Both types of story
aimed to explain and illustrate a different way of life when community bonds were stronger and people
behaved better, despite having fewer material possessions and less economic security. Most of the hardship
stories concerned insufficiency in food or clothing or harsh working conditions, but Lillian’s story of her
first day at work differs from the majority since it addressed the social and moral consequences of such
hardship.
An alternative approach would be to focus on the form or structure of the story as an example of a
particular type of talk and social interaction. In so doing, we might explore the similarities in type between a
told story and the classical story-forms (such as epic, tragic, comic, romantic, etc.). We might examine how
the narrative progresses: do things get better, worse or remain much the same? We might also look for
evidence of cultural resources in the language used or meanings attributed to the words. Such questions have
been applied to the narratives of chronic illness discussed earlier (Bury, 2001).
In Language in the Inner City (1972), Labov identified five elements that he argued were present in all
narratives. In addition to these five elements, a narrative might be preceded by an abstract alerting the
audience to the nature and meaning of the story to be related. Whether or not an abstract is included, all
stories begin with an orientation outlining who is involved in the story and when and where the event takes
place. This is followed by the description of a complicating action that is the core of the narrative, followed
by an evaluation of the significance of what has happened and some form of resolution before a coda
signifying that the narrator has relinquished the conversational lead.
As summarised in Figure 21.1, Lillian’s story fits well with this outline structure with one important
exception. The exception is that Lillian’s story contains no apparent resolution that might explain why she
could recollect it so clearly after a passage of 62 years. In order to understand the story better, we have to
look at its place within a broader narrative comprising of two elements. The first element was Lillian’s own
sense of identity. The second element was the narrative co-produced by Lillian and her husband George
concerning their respective childhoods and whether people living in their locality in the 1930s and 1940s
had ‘felt’ poor since everyone was similarly deprived.

Figure 21.1 An illustration of the application of Labov’s narrative structure to Lillian’s story of her first day
at work
 Labov’s narrative structure (1972) Elements in Lillian’s story of her first day at work
1. Abstract ‘The only time I felt poor …’
2. Orientation Wartime, evacuation, leaving school at 13, Lillian’s mother getting her a job. Smith’s
the tailors as a shop and a social environment.
3. Complicating action Lillian arrives in a home-made crêpe dress, embarrassment.
4. Evaluation How the event made Lillian feel, her mother’s lack of understanding
5. Resolution ??
6. Coda ‘Cos that’s how life was in those days ...’
The lead-up to Lillian’s story of her first day of work, reproduced below, provides further insight into
the meaning of this story and its effects on the conversation that followed it:

LILLIAN: But we were all like that.

GEORGE: Oh yes, yes, we were.

LILLIAN: This is what I’m trying to say. We weren’t really poor because
we were all like it. We were fed, well, country fed, you know. We were filled. ... But
talking like that, I didn’t feel poor because everyone was poor. The only time I felt
poor and I did feel this ...

In the act of remembering and recounting the story, Lillian changes her stance towards poverty from
something that was hardly felt, to a source of great shame. As she relinquishes the conversational lead at the
end (Labov’s ‘coda’), she prompts George to talk about his experiences at a school attended by children
from much wealthier families. This reiterates the theme of class divide. Towards the end of the interview,
George voiced the opinion that if given the choice between going forward or going back 80 years, he would
go back ‘like a shot’. However, Lillian disagreed:

I think people were nicer in those times. They helped each other. But I don’t think I
personally could live in [those times]. I’m talking about now what I see of London
and that, and how this area was. I don’t like squalor. That hurts me how I read how X
was. If you read that book on X, it’s disgusting and I don’t like that. I don’t think I
could have lived with that.

Lillian’s description of her reaction to a book containing photographs and descriptions of physical poverty is
contradictory. She says that she does not think that she could live in conditions of squalor but the researcher,
George and Lillian herself are all aware that her family was particularly poor even compared with others in a
deprived area. The significance of the story of the first day at work is that it marked a transition from a state
of unaware deprivation to an acute consciousness of difference. That process was, however, incomplete and
it is in Lillian’s later life that the enormity of having been connected with squalor impinges on her moral
consciousness: ‘That hurts me …, it’s disgusting’.
 The significance of the story of the first day at work occurs at different levels. It provides an example of
lived experience of social marginalisation. It illustrates how ordinary people tell stories in similar ways to
great works of fiction and in so doing reveal dimensions of meaning and interpretation that other research
methods may not access. Finally, it illustrates that the very act of story-telling may change the direction of
an account as the people once again re-evaluate what has happened and seek some form of resolution. If
narratives that relate to sensitivities concerning personal identity are often rehearsed, it may be because no
single telling can provide a complete account of something so important and complex.

21.5.5 Validity and Trustworthiness

Riessman (1993) argues that the starting point for evaluating the credibility of a narrative analysis must be
an explicit acknowledgement by the researcher that their analysis is the production of particular discourses
or theoretical frameworks. This enables the researcher first of all to dismiss criteria for validity based on
realist assumptions, and, second, to acknowledge that a different theoretical framework might produce a
different analysis. Therefore, as Riessman points out, the basis for assessing the validity of an analysis no
longer resides with the impossible task of representing the ‘truth’ but instead focuses on the notion of
‘trustworthiness’. An analysis should not claim to be any more ‘truthful’ than another, but rather render
transparent the process by which the interpretation of the narrative and stories has been reached. Then we
can argue that there is a high degree of trustworthiness in the analysis and any conclusions drawn from it.

21.6 Summary

This chapter has provided an introduction to both the theoretical assumptions and the practical
methodological issues concerned with narrative analysis. Although there is no single answer to questions
such as ‘What is a narrative?’ and ‘How do we analyse them?’, there are a number of shared assumptions
that will help guide our analysis and which can deepen our understanding of social research as a social
activity.
As discussed throughout this chapter, narrative research is concerned not only with the content of a
narrative but also the social context in which stories are told and the influence of different groups
(Plummer’s ‘producers’, ‘coachers’/’coaxers’, and ‘consumers’) on the act of narration. Stories that are
narrated publicly (which may include a research interview) are often concerned with the individual’s sense
of self and are used to establish or negate an aspect of identity. People often tell stories or talk in a storied
form about experiences that are socially stigmatised, traumatic or unresolved and so narrative research can
offer insights into deeply held cultural values and assumptions. Although this chapter has focused mainly on
the analysis of narratives present in interview data, the principles presented apply equally well to narratives
identified in textual material, focus group interactions and other types of data. Different research questions
will influence decisions about the unit of analysis (categorical or holistic) and its focus (content or form).
However, whatever our research question and focus, the theoretical assumptions underpinning narrative
analysis will challenge the more traditional realist approach to data collection. This in turn renders
problematic the realist assumptions of reliability that are often used to assess the validity of a researcher’s
work and we will need to find other ways to render transparent the process of analysis and establish the
trustworthiness of the analysis presented.

DISCUSSION QUESTIONS
1. What does narrative analysis mean?
2. What do you think about the challenge narrative analysis poses to the traditional assumptions
underpinning the collection of qualitative data?
3. What can be gained from focusing on the social production of the story?
4. What are the advantages and disadvantages of carrying out interviews designed to elicit
narratives in comparison with the unplanned generation of narrative-like data during qualitative
interviews?
5. What is the basis for judging ‘good’ narrative research?

PROJECT

You have been asked to carry out narrative interviews on the subject of childhood with three people of
different ages. Ideally, at least one of the interviewees should be aged 70+ and no interviewee should be
younger than 25. The interviews should last approximately one hour. You need to develop an interview
guide consisting of six or seven main questions together with subsidiary prompts. One of the interviewees
should be interviewed on two occasions so that you can explore the advantages (and any disadvantages)
of multiple interviews with the same person.

RESOURCES
Plummer (1995) Telling Sexual Stories provides an excellent introduction to a sociology of story-telling.

Riessman (1993) Narrative Analysis and Lieblich et al. (1998) Narrative Research: Reading, Analysis, and
Interpretation also offer clear introductions to the nature of narrative research and the different techniques
used in narrative enquiry.

The collection edited by Andrews et al. (2000), entitled Lines of Narrative: Psychosocial Perspectives,
offers examples of studies that have used narrative approaches, either in isolation or in combination with
other methods.
Articles by Atkinson (1997a) and Frank (2000) in the journal Qualitative Health Research provide a lively
entry into the debates surrounding narrative research.